By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Errata: Corrigenda Volume 61, Issue 5, 581, Article first published online: 14 September 2007
Susana Jiménez-Murcia, PhD, Department of Psychiatry, University Hospital of Bellvitge c/Feixa Llarga, s/n, 08907-Barcelona, Spain. Email: firstname.lastname@example.org
Abstract The aim of the present study was to determine whether anorexia nervosa (AN), bulimia nervosa (BN) and obsessive-compulsive disorder (OCD) share clinical and psychopathological traits. The sample consisted of 90 female patients (30 OCD; 30 AN; 30 BN), who had been consecutively referred to the Department of Psychiatry, University Hospital of Bellvitge, Barcelona. All subjects met DSM-IV criteria for those pathologies. The assessment consisted of the Maudsley Obsessive-Compulsive Inventory (MOCI), Questionnaire of obsessive traits and personality by Vallejo, Eating Attitudes Test-40 (EAT-40), Eating Disorder Inventory (EDI), and Beck Depression Inventory (BDI). ancova tests (adjusted for age and body mass index) and multiple linear regression models based on obsessive-compulsiveness, obsessive personality traits and perfectionism, as independent variables, were applied to determine the best predictors of eating disorder severity. On ancova several significant differences were found between obsessive-compulsive and eating-disordered patients (MOCI, P < 0.001; EAT, P < 0.001; EDI, P < 0.001), whereas some obsessive personality traits were not eating disorder specific. A total of 16.7% OCD patients presented a comorbid eating disorder, whereas 3.3% eating disorders patients had an OCD diagnosis. In the eating disorder group, the presence of OC symptomatology was positively associated (r = 0.57, P < 0.001) with the severity of the eating disorder. The results were maintained after adjusting for comorbidity. Although some obsessive-compulsive and eating disorder patients share common traits (e.g. some personality traits especially between OCD and AN), both disorders seem to be clinically and psychopathologically different.
The elevated comorbidity between obsessive-compulsive disorder (OCD) and eating disorders (ED), together with phenomenological as well as clinical similarities have led numerous authors to suggest common etiopathogenic roots that would fall in the obsessive-compulsive spectrum.1–7
Evidences for a relationship between OCD and ED can be found in epidemiological data on comorbidity: the lifetime prevalence of OCD among subjects with ED varies between 11% and 41%,8–10 whereas in OCD patients the prevalence of ED ranges between 11 and 13%.11,12
Regarding personality traits, high levels of obsessionality have been found both in OCD and ED,5,13,14 even as a risk factor for the development of ED later in life.14,15 Even though the restricted type of anorexia nervosa (AN) has generally been associated with perfectionism, preoccupations with cleaning and high levels of rigidity and persistency,16,17 both in male and female subjects,18 some of these traits has also beendescribed in bulimia nervosa (BN).19,20 Likewise, a similar prevalence of personality disorders (20–25% of cases), particularly cluster C,21,22 has been observed in both disorders.
Only few studies in the literature have focused on the phenomenology of OCD versus ED. In comparison to OCD, obsessions and compulsions related to symmetry and order23 have predominantly been described in ED patients. In addition, other studies have revealed clearly different phenomenological traits between both pathologies (e.g. egodysthonia vs egosynthonia).24
Regarding familiar and individual psychopathology, the presence of obsessive-compulsive personality in first-degree relatives15,25 and an OCD during childhood9 have been identified as risk factors for the development of ED later in life.
At a genetic molecular biochemical level, the results are still inconclusive. The neurotransmitters most frequently studied in both disorders have been those involved in the serotonergic system.7,26 In this regard, whereas some studies report a positive association with AN,27 even when compared with OCD,28,29 others find no relationship.30 Other family studies found similar genetic transmission in both disorders.25,31 This shared biological vulnerability has also been postulated from a biochemical point of view, both in ED32 and OCD.33
In relation to clinical variables and therapy effectiveness, several ED studies outline that comorbid OCD is associated with higher eating severity,34 longer duration of the disorder,35 and worse prognosis,36,37 whereas other did not.38 Additionally, other studies have associated poorer outcome in ED to premorbid obsessive personality traits,13,37 which seem to persist even after recovering from an ED.38,39
Therefore, despite some empiric evidence suggesting an association between OCD and ED, the results are contradictory. To our knowledge the relationship between both disorders at a symptomatological level, while controlling for the variables age and weight, has rarely been investigated.
The objectives of the present study were (i) to determine the rate of comorbidity between OCD and ED; (ii) to explore differences in obsessive and eating symptomatology, but also on the frequency of obsessive-compulsive personality traits, between OCD and ED patients by controlling for age and weight; and (iii) to analyze the association between eating symptomatology and obsessionality in AN and BN.
The sample consisted of 90 female patients (30 OCD, 30 AN and 30 BN) diagnosed according to DSM-IV40 criteria and consecutively referred for assessment and treatment to the Department of Psychiatry, University Hospital of Bellvitge, Barcelona. The mean patient age was 25.56 ± 7.14 years, mean age of onset of the disorder was 17.7 ± 5.0 years, and the mean duration of the disease was 9.2 ± 7.1 years.
Exclusion criteria were a past or present history of substance abuse, age under 17 or over 50 years, and male gender (ED, 8.2% male; OCD, 51.5% male). In the ED group, specifically, a diagnosis of eating disorder not otherwise specified (40% of the original ED recruited sample) was also considered as an exclusion criterion.
Concerning subtypes of disorder, 85% of BN patients had the purging subtype, and 64% of AN patients had the restrictive subtype. Among OCD patients, 58% showed checking compulsions, 28% exhibited cleaning compulsions, and 14% had sexual-content obsessions.
The patients were given the Maudsley Obsessional-Compulsive Inventory (MOCI)41 (validated Spanish version of the MOCI42,43), the Questionnaire of Obsessive Traits and Personality by Vallejo (CRPO)44 (validated Spanish version of the CRPO44,45), the Eating Attitudes Test (EAT-40)46 (validated Spanish version of the EAT-4047), the Eating Disorder Inventory (EDI)48 (validated Spanish version of the EDI49) and the Beck Depression Inventory (BDI)50 (validated Spanish version of the BDI51).
All participants underwent medical–physical and psychopathological assessments before starting treatment. The OCD and ED patients were diagnosed by means of the Structural Clinical Interview for DSM-IV Axis-I mental disorders (SCID-I).52,53 A battery consisting of the previously mentioned self-reported psychometric tests was administered to all patients. The Ethics Committee of University Hospital of Bellvitge, Barcelona approved the present study and written informed consent was obtained from all participants.
SPSS 12.0 software (SPSS, Chicago, IL, USA) was used for statistical analysis. Because the three groups showed significant differences in age and body mass index [BMI = weight (m)/Height2 (m)2], anova and ancova tests were carried out in order to compare each dependent quantitative variable with regard to diagnostic group by using age and BMI as covariates. When significant differences were obtained, multiple comparison tests (Scheffé test) were applied. Furthermore, in both ED groups, multiple linear regression models (stepwise) were carried out in order to determine which factors of the obsessionality domain (as measured by the MOCI, the CRPO and the EDI [Perfectionism subscale]) could best predict ED severity, as measured by the EDI total score. Given that comorbid OCD and ED was present in some cases, we adjusted the analyses for this variable a posteriori in order to dismiss its effect. A significance level of 0.05 was considered acceptable.
As expected, significant differences between the AN group and the remaining two groups (BN and OCD) were observed with respect to weight (F = 33.7; d.f. = 2; P < 0.001) and BMI (F = 44.03; d.f. = 2; P < 0.001). Similarly, significant age differences were detected between groups (F = 6.47; d.f. = 2; P < 0.014), in that OCD was the older group (mean age, 29.2 years), compared to both ED (AN and BN: 23.6 years and 23.9 years, respectively).
Psychometric comparison by group
As shown in Table 1, OCD and ED patients showed significant mean differences in negative attitudes towards eating and eating psychopathology (total EDI scores), with OCD patient scores falling in the normal range.54 On anova significant differences were seen between groups in all EDI subscales, excepting for Perfectionism (P < 0.133) and Maturity Fears (P < 0.055). Post-hoc tests (Scheffé test) showed that ED patients scored significantly higher than OCD patients in the following subscales: Drive for Thinness (P < 0.001), Interoceptive Awareness (P < 0.001), and Ineffectiveness (P < 0.001). Furthermore, BN patients had higher values in Bulimic Symptomatology (P < 0.001) and Body Dissatisfaction (P < 0.001) subscales in relation to the other two clinical groups (AN and OCD). Most of these results were maintained after adjustment for age and BMI. However, when including comorbidity as a covariate, statistical differences appeared regarding the EDI-Perfectionism scale, with AN patients scoring significantly higher than OCD patients (P = 0.019).
Table 1. Group comparison of eating habits and symptomatology
OCD (n = 30)
AN (n = 30)
BN (n = 30)
F (d.f. 2.87)
AN, anorexia nervosa; BN, bulimia nervosa; EAT-40, 40-item Eating Attitudes Test; EDI, Eating Disorders Inventory; EDI subscales: EDI-BD, Body Dissatisfaction; EDI-BUL, Bulimia; EDI-DT, Drive for Thinness; EDI-I, Ineffectiveness; EDI-IA, Interceptive Awareness; EDI-ID, Interpersonal Distrust; EDI-MF, Maturity Fears; EDI-P, Perfectionism; OCD, obsessive-compulsive disorder.
As shown in Table 2, with respect to OC symptomatology, on anova significantly higher MOCI and CRPO mean scores were seen in the OCD group in comparison to ED patients (P < 0.001). MOCI Checking (P < 0.001) and Cleaning (P < 0.019) subscales yielded the clearest differences between OCD and ED patients. Both groups also differed significantly in the level of depressive symptomatology, with ED patients having higher BDI scores (P = 0.001).
Table 2. Group comparison with respect to obsessive-compulsive symptomatology (MOCI), obsessive personality (CRPO) and depressive symptoms (BDI)
OCD (n = 30)
AN (n = 30)
BN (n = 30)
F (d.f. 2.87)
AN, anorexia nervosa; BN, bulimia nervosa; BDI, Beck Depression Inventory; CRPO, Questionnaire of Obsessive Traits and Personality; MOCI, Maudsley Obsessive-Compulsive Inventory; MOCI subscales: MOCI CD, Consciousness/Doubt; MOCI-Ch, Checking; MOCI Cl, Cleaning; MOCI SR, Slowness/Repetition; OCD, obsessive-compulsive disorder.
Because 40.9% of OCD patients and 20.7% of BN patients were overweight (BMI > 25), and considering that age was significantly different between groups, we re-analyzed our data with ancova using age and BMI as covariates. These analyses confirmed the previous results, excepting for the CRPO scale, which lost significance (F = 1.65; d.f. = 4; P < 0.173). The adjustment for comorbidity also maintained these differences at a non-significant level.
Comorbidity between OCD and ED
Five patients in the OCD group (16.7%) presented a comorbid ED. Three of those fulfilled criteria for an eating disorder not otherwise specified (EDNOS), one for purging BN and one for binge eating disorder (BED). Among ED patients, two (3.3%) met DSM-IV criteria for OCD, both of them in the AN group.
Association between obsessionality and ED
Table 3 shows the results of linear regression models measuring the effect of obsessionality on severity of ED symptoms, both in AN and BN groups. The best predictive model was that for the BN group, which explained 50.1% of the total variance in ED symptoms. According to these results, higher MOCI scores (B = 2.09; 95% confidence interval [CI]: 0.70–2.01; P = 0.005) and EDI-Perfectionism scores (B = 2.66; 95%CI: 0.83–4.49; P = 0.006) predicted higher EDI scores. Then, severity of ED symptomatology increases as OC symptomatology increases. These results were maintained after adjusting for comorbidity with OCD. A positive correlation was observed between total EDI score and total MOCI score (r = 0.574; P < 0.001) in the total ED sample (Fig. 1).
Table 3. Effect of OC symptoms and traits on severity of eating symptomatology (EDI total score) in AN and BN
95%CI for B
Only statistically significant results are shown. Dependent variable: EDI total score.
The aim of the present study was to explore differences in eating symptomatology and OC traits between ED and OCD patients, comorbidity between both disorders, and the effect of obsessionality in eating symptomatology.
Regarding psychometric variables, OCD patients had lower scores on eating disorder scales than ED patients, even after controlling for BMI, age, and comorbidity with ED. These differences were clear with respect to traits such as interoceptive awareness and ineffectiveness. Some authors suggest in this regard that these traits are especially characteristic of ED.34,55
The OCD group presented more obsessive symptomatology than ED patients, even after controlling for BMI, age, and comorbidity. Although these results are consistent with those from other studies,34,35 other authors do not find these differences.56 This inconsistency might be related to the type of sample used. Whereas in the Halmi et al. study AN patients with obsessive symptomatology did not differ from OCD patients either in relation to symmetry and somatic obsessions or in compulsions of order and hoarding,56 the present OCD group consisted of 58% of patients with checking rituals, 28% with cleaning rituals and 14% with sexual-content obsessions.
Concerning the shared comorbidity between OCD and ED, 16.7% of OCD patients presented a comorbid ED, whereas only 3% of ED patients exhibited a comorbid OCD (all of them from the AN group: 6.7%). This result finds support in some studies,8,35,55,57,58 but not in others.19,34
The OCD and ED patients exhibited high levels of obsessive personality, even after controlling for BMI, age, and comorbidity. This is consistent with other studies that found marked obsessive personality traits in AN and OCD.5,37,39
Both AN and BN patients had higher levels of depressive symptomatology than OCD patients, who exhibited slight to moderate symptomatology. This result suggests that ED patients suffer stronger emotional disturbances (depressive symptoms) than do OCD patients. Several authors have found similar results and highlight the importance of these symptoms for both disorders.59 Also, in concordance with other studies, the level of obsessionality was positively associated with the severity of eating symptomatology.16,35
In conclusion, the present study suggests that OCD and ED are different from a clinical and psychopathological point of view. Nevertheless, similarities appeared at the level of personality in a subgroup of OCD and AN patients. This association suggests the possible existence of a similar vulnerability diathesis in some cases. Regarding the therapeutic implications of the present findings, basic components of cognitive behavioral therapy in this subgroup of patients should be especially focused on cognitive dysfunctional styles (e.g. rigidity, black–white thinking), which seem to act as one of the most important maintenance factors of symptomatology.
The limitations of the present study include: (i) small sample size, which meant that comparisons between subgroups of patients were unable to be done; (ii) lack of consideration of the influence of variables such as anxiety level, other personality traits, presence of comorbid personality disorders etc., and the effect of other variables such as phenomenological, genetic, and familiar psychiatric history; and (iii) lack of standard and more comprehensive assessment tools for obsessive-compulsive symptoms.
Future research in this field is needed in order to determine and delimit this subgroup of ED patients with high level of obsessive personality, and to establish whether these patients also generally exhibit the higher egosynthonia and lower motivation for treatment that are associated with poorer response to treatment and worse prognosis.
Financial support was received from Instituto de Salut Carlos III (CIBER, CB/06/03/0034; PI051336) and the Generalitat de Catalunya (2005SGR322). Isabel Krug was recipient of an FI fellowship from the Generalitat de Catalunya (2005FI425).